wnoise

49 ms·

541.
by wnoise 7y ago
Yes, saddle points are far more common than local minima in high dimension. Unfortunately they're really good at slowing down naive gradient descent...
542.
by wnoise 7y ago
This general statement is far short of currently being a general norm. But his argument stands if it is being pushed as a norm , with attempts to either formally enshrine this as a norm or policies that make it easier to enforce as a norm.
543.
by wnoise 7y ago
> you politely believe them and move on, comfortable in the knowledge that it may or may not be true. You have a very different understanding of the word "believe" than I do.
544.
by wnoise 7y ago
Many of them are not government organizations. Even the ones that are, are state organizations, not federal. They do all take federal money though...
545.
by wnoise 7y ago
He'd have a very hard time winning. In many jurisdictions there is a "substantial truth" doctrine: if a statement gives an impression that's substantially close to the truth, it's considered OK, even if technically
546.
by wnoise 7y ago
Right, because you can't answer that without prior information. Log-odds is a step towards giving a function of prior information (but it is limited to the actual class of hypotheses investigated).
547.
by wnoise 7y ago
> Interactions between particles, waves, fields, etc. are quantized, Yes. > which means there is a set of distances at which a given interaction can happen, and a set of distances at which it can't. ... that's not how quanti
548.
by wnoise 7y ago
Well, you can report odds ratios (or their log).
549.
by wnoise 7y ago
Both can be true, but if society has already spent a lot on one, than it can be much much more efficient to gather the low hanging fruit on the other.
550.
by wnoise 7y ago
Making lots of one-way streets is certainly an overused hammer, but it shouldn't double the distance traveled, it should add at most three blocks, and usually far less.
551.
by wnoise 7y ago
In the sense that it ignores that coherence. Any state on the disc of the Bloch sphere's equator has <z> = 0, and 50:50 chance of measuring 1 or 0. If you want to assign a real value of 0.5 to a bit, then the only one that make
552.
by wnoise 7y ago
> Also, int foo(void) is illegal in c++. Uh, citation needed. It would make compatibility with C libraries harder for no good reason. I believe it is valid as both declaration and definition. Both g++ and clang++ accept it without even
553.
by wnoise 7y ago
Facebook is millions , if not billions of tiny public forums. The problem is that people seem to want them all to have the same standards. If you don't want stuff like this on your wall or feed, it's easy to keep out. If the a
554.
by wnoise 7y ago
Life often incentivizes immoral actions. That's not usually considered a mitigating factor.
555.
by wnoise 7y ago
Um sure, I guess. I really would like to understand how and why you disagree though, but I think if I said anymore, it would just be reiterating my points in a not terribly useful way.
556.
by wnoise 7y ago
What you have quoted is just talking. It isn't something that needs any strong worry about consent. Read standalone, it could be anything from mutually flirtatious banter to the cringiest of escalation attempts. Read in the context
557.
by wnoise 7y ago
Many people would consider being nice, friendly, and especially charming to be flirting.
558.
by wnoise 7y ago
> Men used to be able to basically hit on women in all contexts with zero consequence Zero legal consequence -- social consequence or vigilante consequence was always a possibility. > More recently, people have been pushing back on me
559.
by wnoise 7y ago
Well, FOSTA/SESTA has already started to dismantle it...
560.
by wnoise 8y ago
> - object memory layout and API are unnecessarily bound together, so you can't change so much as a single function argument without completely recompiling all dependent code. These can readily be put in different headers. People o
561.
by wnoise 8y ago
> There was democratic support for slavery in the US at one point, something to keep in mind. Only if you don't count the votes of the slaves, of course.
562.
by wnoise 8y ago
> because most of the CO2 stays in the body Say what? ... not at all. No CO2 is "stored", though it may temporarily hang out in buffers such as the blood. But they have a very limited capacity and the flow out is extremely cl
563.
by wnoise 8y ago
They do move in tandem on a molar basis, but because there is so much more O2 in the air, even very high concentrations of CO2 have barely changed the concentration of O2.
564.
by wnoise 8y ago
It was my understanding that the original Maxwell equations used separate letters for each component rather than indices or any modern method of writing down the components as a unified thing (so the four expand to at least 12 separate equa
565.
by wnoise 8y ago
> why did they pick Clojure to do probabilistic programming? As most language decisions for new projects, it probably came down to "it's their favorite language". > btw, Python programmers have the option of pymc3 htt
566.
by wnoise 8y ago
The rapid rise of self-diagnosed "gluten intolerance" for instance.
567.
by wnoise 8y ago
It's awful precisely when it would start delivering a different result than FPTP, because that's when strategic voting comes into play.
568.
by wnoise 8y ago
COINTELPRO wasn't public knowledge at the time, and many massive surveillance systems operating recently and today are known, e.g. PRISM.
569.
by wnoise 8y ago
Yes, they are, which is as much as argument that religious beliefs are too protected as it is an argument that political beliefs are not protected enough.
570.
by wnoise 8y ago
Montana is the only exception.

More ›